EM algorithms for sequence estimation over random ISI channels

This paper presents a new approach using EM (expectation-maximization) algorithms for ML (maximum likelihood) sequence estimation over unknown ISI (inter-symbol interference) channels with random channel coefficients. By using the EM formulation to marginalize over the channel coefficient distribution, maximum-likelihood estimates of the transmitted sequence are obtained. The EM algorithms are shown to perform better, in terms of BER, than existing algorithms which perform jointly-optimal sequence and channel estimation.